Currently, androgen ablation is the most effective therapy available for advanced prostate cancers. However, androgen ablation does not prevent prostate cancer progression to the lethal androgen-independent (AI) state. The goal of this proposal is to develop a gene therapy strategy for AI prostate cancer based on our study of the regulation of prostate specific antigen (PSA) and prostate specific membrane antigen (PSMA) promoter/enhancer in AI prostate cancer cells. PSA expression is upregulated by androgens. In contract PSMA expression is down regulated by androgens. We believe that PSA and PSMA promoters likely function weakly in patients receiving androgen-ablation therapy due to the presence of low levels of androgen and AR mutation or amplification that will partially activate PSA promoter and suppress PSMA promoter. In the previous funding period, by detecting and deleting silencers, we identified strong central enhancer cores from PSA and PSMA enhancers, called AREc3 and PSME(del2). A combination of AREc3 and PSME(del2), called PSES, demonstrated strong prostate-specific activity regardless of androgen status. This observation suggests that AREc3 and PSME(del2) can work together and compensate each other's activity in patients with AI prostate cancer. We propose to investigate the tissue-specific activity of AREc3 and PSME(del2) when placed adjacent to each other in adenoviral vectors. We constructed a recombinant adenovirus, Ad-PSES-Luc, carrying luciferase reporter under the control of PSES. We will test its tissue-specific activity in PSA/PSMA-positive prostate cancer cells, PSA/PSMA-negative prostate cancer cells, non-prostatic cancer cells and primary human tissue cultures from different organs. We hypothesize that AREc3/PSME(del2) or PSES can control adenovirus replication in a prostate cancer-specific manner by controlling adenoviral E1a and E1b expression. We will construct several AREc3/PSME(del2) or PSES based-prostate-restricted replicative adenoviruses (PRRAs) to test this hypothesis. The majority of prostate cancers compose both PSA/PSMA positive and negative cells that might limit the therapeutic efficacy of our PRRA. To overcome this potential hurdle, we propose to incorporate a tumor specific apoptosis inducer, Trail, into our PRRA. We hypothesize that Trail can eliminate neighboring PSA/PSMA negative tumor cells when expressed by PSA/PSMA positive tumor cells. We will generate Trail-armed adenovirus and a chimeric tumor model, composing both PSA/PSMA positive and negative tumor cells, to test this hypothesis. We will potentially provide a novel agent to treat androgen-independent prostate cancer at the end of this project.